JP4637577B2 - Method for transmitting an audio signal according to a pixel transmission method for determining priority - Google Patents

Abstract

A method for the transmission of audio signals between a transmitter and at least one receiver operates according to the prioritizing pixel transmission method. The audio signal is first broken down into a number of spectral fractions. The broken-down audio signal is stored in a two-dimensional array with a plurality of fields. The dimensions to be registered in the field are frequency and time; the value to be registered in the field is amplitude. Groups are then formed from the individual fields and a priority is assigned to the individual groups, in which the priority will be gauged as higher if the amplitudes of the group values are higher, and/or if the amplitude differences of the values of one group are higher, and/or if the groups are closer to actual time. Finally, the groups are transmitted to the receiver according to the order of their established priority.

Description

  The present invention relates to a method for transmitting an audio signal according to a pixel transmission method for determining priority, as described in claim 1 of the appended claims.

  Currently, there are many different methods for compressing and transmitting audio signals. In general, the following methods exist.

A method for reducing the sampling rate, eg 3 kHz instead of 44 kHz;
-A method of transmitting the sampling values non-linearly, for example ISDN transmission;
A method of using a pre-stored acoustic sequence, for example MIDI or voice simulation;
A method that uses a Markov model to correct transmission errors.

  The commonality between these known methods is that there is a satisfactory level of speech understanding even when the transmission rate is low. This is largely achieved by averaging. However, if the sampling rate is low, voices of different sources will result in voices that sound the same on the receiver side, so that voice fluctuations that are recognizable, for example in normal conversation, are no longer transmitted. As a result, communication quality is significantly limited.

  A method for compressing and decompressing image data or video data by pixel transmission with a predetermined priority is described in German patent applications DE101138880.6 (corresponding to PCT / DE02 / 00987) and DE10152612.1 (corresponding to PCT / DE02 / 00995). Are listed. In these methods, for example, digital image data or video data consisting of an array of individual picture elements (pixels) is processed, where each pixel describes temporal or temporal information that describes the color or luminance information of the pixel. It has a pixel value that changes to According to the invention, a priority is assigned to each pixel or group of pixels, and each pixel is stored in an array of priorities according to its priority. The array has pixel values sorted according to priority at each time point. Depending on the priority, these pixels and the pixel values used for the priority calculation are also transmitted or stored. A pixel gets a high priority when the difference to its neighboring pixels is very large. To reconstruct, each actual pixel value is displayed on the display. Pixels that have not yet been transmitted are calculated from the pixels that have already been transmitted. This method is also used in principle for the transmission of audio signals.

  It is an object of the present invention to provide a method for transmitting an audio signal that operates with as little loss as possible even when the transmission bandwidth is narrow.

  This problem is solved according to the invention by the features of claim 1 as claimed.

  According to the present invention, the audio signal is first decomposed into n spectral components. The decomposed audio signal is stored in a two-dimensional array with a plurality of fields, which have frequency and time as dimensions, and amplitude as the value to be entered in that field. A group is then formed from each individual field of the array and at least two fields adjacent to that field, and a priority is assigned to each group, where the priority of the group is the amplitude of the group value, respectively. Is larger and / or the amplitude difference between the values of the respective groups is larger and / or the closer the group is to the actual time, the higher is selected. Groups are then transmitted to the receiver in order of their priority.

This new method is largely based on Shannon's theorem . Therefore, if the signal is sampled at twice the frequency, the signal is transmitted without loss. This means that the sound can be resolved into individual sine vibrations of different amplitudes and frequencies. Thus, the acoustic signal is unambiguous by the transmission of individual frequency components including amplitude and phase, and is formed again without loss. In that case, in particular, frequently used sound sources such as musical instruments and human voices are composed of resonators, and the resonance frequency does not change or changes only slightly.

  Preferred embodiments and further developments of the invention are described in the dependent claims.

  Embodiments of the present invention will be described below. In that case, reference is also made, in particular, to the specifications and drawings of the earlier patent applications DE 10113880.6 and DE 10152612.1.

  First, sound is recorded, converted into an electrical signal, and decomposed into its frequency components. This can be done by FFT (Fast Fourier Transform) or by n individual frequency selective filters. If n individual filters are used, each filter absorbs only a single frequency or a narrow frequency band (similar to cilia in the human ear). Therefore, at each time point, the frequency and the amplitude at this frequency are obtained. In this case, the number n can take different values depending on the terminal device characteristics. The larger n is, the better the audio signal can be reproduced. Therefore, n is a parameter that can scale the quality of audio transmission.

  This amplitude value is temporarily stored in each field of the two-dimensional array.

  In that case, the first dimension of the array corresponds to the time axis and the second dimension corresponds to the frequency. Thus, each sampling value having a particular amplitude value and phase is uniquely defined and can be stored as an imaginary number in the corresponding field of the array. Thus, the audio signal is represented in the array with three acoustic dimensions (parameters). That is, the first dimension of the array, for example, a time in milliseconds (ms) and perceptually distinguishable as a duration, and the frequency of hertz (Hz) as the second dimension of the array. It is perceptually recognized as the tone height and the energy (or intensity) of the signal stored as a numerical value in the corresponding field of the array, and perceptually as the volume or intensity of the sound. It can be distinguished.

  In the comparison between the application DE10113880.6 and the DE10152612.1, for example, the frequency of the audio signal corresponds to the image height, the time corresponds to the image width, and the amplitude (intensity) corresponds to the color value.

  Similar to the method of prioritizing pixel groups in image / video coding, each group is formed from adjacent values and prioritized. Each field is considered per se and forms a group with at least one, but preferably a plurality of adjacent fields. Each group consists of a position value determined by time and frequency, an amplitude value at that position value, and an amplitude value of peripheral values according to a predetermined format (see FIG. 2 of the applications DE10113880.6 and DE1015272.1). reference). In that case, in particular, a group that is close to the actual time and / or whose amplitude value is very large compared to other groups and / or whose amplitude values are significantly different from each other within the group is very high priority. Get ranking. Pixel group values are sorted in descending order and stored or transmitted in this order.

  The width (time axis) of the array preferably has only a limited spread (eg 5 seconds), ie, only signal portions that are, for example, 5 seconds long are always processed. After this time (eg 5 seconds), the array is filled with each value of the subsequent signal part.

  In accordance with the prioritization parameters described above (amplitudes, temporally close values, and amplitude differences with respect to adjacent values), the values of the individual groups are received in the receiver.

  At the receiver, the group is again entered into the corresponding array. According to the patent applications DE10113880.6 and DE10152612.1, a three-dimensional spectral display can then be formed again from the transmitted group. The more groups each are received, the more accurate the reconstruction. Array values that have not yet been transmitted are calculated from the already transmitted array values using interpolation. From the array thus formed, a corresponding audio signal is then generated in the receiver, which can then be converted to sound.

  For example, n frequency generators can be used to synthesize the audio signal, which are added to the output signal. This parallel structure of n generators gives good scalability. Furthermore, since the clock rate is significantly reduced by parallel processing, the playback time of the mobile terminal is increased with less energy consumption. For parallel use, for example, a simple structure FPGA or ASIC can be utilized.

  The method described above is not limited to audio signals. In particular, the method uses a plurality of sensors (sound sensors, light sensors, tactile sensors, etc.) that continuously measure signals and then display the signals in an (nth order) array. Wherever possible, it can be used effectively anywhere.

  The advantage over conventional systems is that they can be used flexibly even in the case of high compression ratios. By utilizing arrays supplied from different sources, synchronization of the different sources is automatically obtained. Appropriate synchronization must be ensured by special protocols or measures in the conventional method. Especially in video transmissions with a long arrival time, for example in satellite connections where sound and image are transmitted over different channels, lip synchronization errors to sound often occur. It can be removed to some extent by the method described above.

  In audio, image and video transmission, the same basic principle of prioritized pixel group transmission can be utilized, so that significant synergies can be exploited during implementation. Furthermore, in this way, simple synchronization is performed between the sound and the image. Furthermore, it can be arbitrarily scaled between image resolution and audio resolution.

  Considering individual audio transmissions based on this new method, more natural reproduction can be obtained in speech. This is because the frequency components (frequency groups) typical for each person are transmitted with the highest priority and thus without loss.

The method and the like for transmitting an audio signal according to the present invention are preferably used for transmitting an audio signal and the like together with an image signal and the like.

Claims (6)

In a method for transmitting an audio signal between a transmitter and at least one receiver according to a prioritized pixel transmission method,
a) decomposing an audio signal into n spectral components by n frequency filters ;
b) storing the decomposed audio signal in a two-dimensional array having a plurality of fields, having frequency and time as dimensions of the two-dimensional array , and further, amplitude as a value to be entered in each field And having
c) by combining the individual fields of the plurality of fields, comprising the steps a field group composed of a plurality of fields, each field group to form a plurality of ones of at least three adjacent field,
d) Prioritizing individual groups of the plurality of field groups . In this case, the priorities of the individual groups are set such that the amplitude of the group value increases and / or the respective groups in the respective groups. The step of increasing the amplitude difference of the values and / or the closer each group is to the actual time,
e) classifying each field group of the two-dimensional array using the priorities;
f) store the field groups in their priority order or in order of priority so that the audio signal is transmitted in a narrow transmission band in order to minimize transmission loss. Transmitting the audio signal to the at least one receiver .   The method of claim 1, wherein the entire audio signal exists as an audio data file and is processed and transmitted as a whole.   The method of claim 1, wherein only a portion of the audio signal is processed and transmitted.   4. A method as claimed in any preceding claim, wherein the audio signal is decomposed into its spectral components using FFT. Within the receiver, the field group transmitted according to its priority is associated with the corresponding array, in which case values not yet transmitted to the array are calculated by interpolation from the values already present. The method according to claim 1, wherein: 6. The method according to claim 1, wherein an electrical signal is formed from the presence value and the calculated value in the receiver and converted into an audio signal.